• Keine Ergebnisse gefunden

Lipo-oligomers optimized towards enhanced lipopolyplex stability

nm, but a highly decreased zeta potential (between 1.5 and 12 mV), indicating the successful PEGylation reaction. However, introduction of 1.0 eq of the bivalent Cys2 -PEG24-GE11 reagent resulted in partial aggregation. As reported previously, the hydrophobic GE11 peptide is known to promote aggregation within pre-PEGylated polyplexes and enhanced particles sizes in case of 454 [226]. An increase in size was identified for all oligomers containing only 4 Stp units, whereas sizes ranged from 284 nm (454) to >5000 nm in case of oligomers containing Y6 (1173 and 1174). An increase to 6 Stp units (1022 and 1023) did not completely help to overcome this aggregation, however with size decreased to a lower extent in case of cholanic acid bearing polyplexes. Lipopolyplexes formed with oligomers containing 8 Stp units exhibited nanoparticle sizes between 71 and 102 nm, and thereby just an increase of approximately 25 nm during PEGylation with Cys2-PEG24-GE11. Polyplexes formed with cholanic acid containing oligomers tended to maintain smaller particle sizes than oleic acid containing polyplexes. Summarizing these findings, we conclude, that the elevated cationic charge is helping to overcome the aggregation potential of hydrophobic peptides [209, 210] such as GE11 by increased nucleic acid compaction.

Consequently, an EtBr compaction assay was carried out to investigate if increased cationic chain-length comes along with improved cargo compaction. For non-PEGylated lipopolyplexes, a remaining EtBr fluorescence was determined between 8 and 17%, while LPEI, which was used as a control exhibited less than 5 % free EtBr fluorescence. PEGylation with 1.0 eq of the bivalent reagents, independent if Ala or GE11 was spiked, did not significantly influence nucleic acid compaction. No differences between oleic acid and cholanic acid could be observed, but oligomers with 8 Stp units generally tended to mediate a slightly improved pDNA compaction compared to oligomers containing only 4 Stp units.

250 IU heparin was added to the same samples, and nucleic acid compaction after polyanionic stress was determined. Here, all lipopolyplexes mediated higher resistance, leading to a maximum of 60% EtBr fluorescence, than LPEI complexes, which released nucleic acid almost completely (>95%). Interestingly, within this experiment polyplexes with Stp2-Y6 (1173 and 1174) retained 60% of pDNA. All other polyplexes maintained pDNA compaction between 40 and 50%, which was already shown for 454 previously [226]. Generally, no notable differences between polyplexes comprising oleic or cholanic acid as stability inducing domains could be observed.

PEGylation of lipopolyplexes did not negatively influence pDNA compaction after polyanionic stress as observed previously for pre-PEGylated polyplexes [133].

As a next experiment, after polyplex formation, stability within physiological salt concentrations, mediated by PBS solution buffered at pH 7.4, was investigated. Here, non-PEGylated polyplexes formed with oligomers consisting of 6 or less Stp units underwent colloidal aggregation with sizes more than 1000 nm. After three hours, all polyplexes apart from polyplexes formed with oligomers containing 8 Stp units were either dissociated or exhibited a size of several 1000 nm. Oligomers 1175-1180 comprising 8 Stp units increased to a maximum size of 370 nm, however, mostly exhibited a size below 250 nm, indicating that the combination of hydrophobic stabilization via peripheral tyrosines (at least trimers) as well as fatty acids (independent if cholanic or oleic acid was introduced) and 8 Stp units per oligomer significantly increased polyplex stability against colloidal stress. Stability of these non-PEGylated lipopolyplexes was maintained for 24 h. PEGylation with 1.0 eq bivalent reagents resulted in different findings. While all polyplexes PEGylated with Ala underwent no aggregation within 24 h, giving proof of the successful PEGylation of these polyplexes, lipopolyplexes PEGylated with GE11 underwent immediate aggregation, latest after 30 min of incubation with PBS. To point out that the behavior of GE11 is linked to lipopolyplexes solely, the well-established LPEI-PEG2k-GE11 [62, 192] was additionally included in this study, however also underwent aggregation 30 min after addition of PBS. This again points out the special behavior of GE11 within polyplexes. Nevertheless, this GE11 mediated aggregation is only an indicator for a reduced stability compared to polyplexes PEGylated with the less hydrophobic Ala, as LPEI-PEG2k-GE11 is well known to mediate EGFR specific gene transfer in-vivo [192, 237].

In a last experiment evaluating biophysical properties, polyplex stability in the presence of serum was investigated. Therefore, polyplexes were formed and incubated in 90%

FBS. In general, as mentioned previously [226], polyplexes remained stable for 24h.

All lipopolyplexes underwent immediate interaction with serum, displayed in an increased size of around 300 nm. This is most likely due to the formation of a serum corona. This interaction is known to facilitate stability, especially for PEGylated polyplexes, and thereby significantly contributes to their stability over time [206].

However, polyplexes PEGylated with Ala tended to (partially) aggregate to a greater

extent than lipopolyplexes PEGylated with GE11, indicating that lipopolyplexes post-modified with GE11 form a more stable serum corona.

Overall these data suggest, that PEGylation is beneficial for lipopolyplex stability and emphasizes that an increased cationic charge density, is required to overcome GE11 mediated polyplex aggregation and to enhance polyplex stability compared to polyplexes formed with oligomers of 4 or 6 Stp units.

Next, lipopolyplexes were post-modified with 1.0 eq of Cys2-PEG24-Ala/GE11 and gene transfer was investigated on the two EGFR positive cell lines Huh7 and KB. After 24 h of incubation on the cells, only the PEGylated 454/pDNA polyplex mediated notable gene transfer in an EGFR specific manner. All other tested oligomers mediated no gene transfer compared to buffer-treated cells, but also none of the polyplexes mediated cytotoxicity within 24 h on the cells. A reason for this could be the enhanced stability of the lipopolyplexes generated from the novel oligomers as investigated by different stress-inducing stability assays. An increased cationic charge density as well as hydrophobic domains (Y6 or cholanic acid), mediated a significantly enhanced lipopolyplex stability in vitro and thereby could hinder DNA release.

Consequently, cellular internalization was investigated. FACS analysis revealed significant intracellular uptake for all Cy5-labeled pDNA/oligomer complexes. A reduced cellular uptake of the lipopolyplexes PEGylated with 1.0 eq of the bivalent Ala control reagent was observed, pointing out that polyplex shielding hinders unspecific cellular uptake in comparison to unmodified lipopolyplexes. Similar findings have been made previously for PEGylated PEI complexes [164]. At the same time, post-modification with 1.0 eq of the bivalent EGFR targeted GE11 PEGylation reagent facilitated cellular uptake at least as good as non-PEGylated core lipopolyplexes. Ogris et al. previously already pointed out in a PEI based work that EGFR-targeting is only affected to a minor extent by PEGylation [54].

Overall, lipopolyplexes formed with oligomers containing 4 Stp units (454 and 1026) mediated a higher uptake than polyplexes formed with oligomers containing 8 Stp units (1175-1180).

Polyplexes with the highest uptake (454, 1026, 1176, 1177, 1178) were transfected again but within these transfections endosomolytic LPEI and chloroquine was added afterwards to reduce the possible lack of endosomal escape during gene delivery. This experiment was conducted as polyplexes formed with the novel lipo-oligomers were successfully taken up in a EGFR specific manner, but could not mediate gene transfer.

Although chloroquine is also known to facilitate polyplex dissociation intracellularly [238, 239], significantly improved transfection efficacy was only found for polyplexes formed with 454. Endosomal escape could be enhanced for parts of the novel oligomers (1026, 1176), although only in a minor extent. Thereby, it still could be considered that the herein generated lipopolyplexes suffer from their enhanced polyplex stability in comparison to the more labile 454/pDNA polyplexes as all polyplexes were internalized facilitating gene transfer theoretically.

However, polyplexes formed with the newly generated lipo-oligomers might retain their cargo in subsequent steps preventing transformation of the pDNA towards its protein.

This question could only be addressed by time consuming microscopy techniques requiring high expertise and labeling of pDNA as well as the oligomers for determination of cellular trafficking.

In general, for all experiments, an equimolar amount of PEG per cysteine within all oligomers was used. This not only resulted in a two-fold excess of PEG per oligomer (as all oligomers contain two cysteines) but also in a two-fold excess of NPys (due to its bivalency) per cysteine of the oligomer. Although this excess mediated the best gene transfer (data not shown) for 454/pDNA polyplexes PEGylated with these bivalent reagents, the excess of reactive PEGylation reagents might reduce uptake to a certain extent, referring to the previously mentioned PEG dilemma [163]. Sizes of (PEGylated) lipopolyplexes were approximately 100 nm for oligomers containing 8 Stp units (1175-1180), with an enhanced stability in the presence of physiological salt concentrations.

At the same time (post-modified) 454 and 1026 pDNA polyplexes exhibited particle sizes of at least 200 nm. This increased size could at least partially help to compensate cellular uptake reduced by the excess of PEGylation reagents as larger particles might be taken up more efficiently [240, 241]. Ogris and colleagues [202] previously showed that polyplex size also plays a critical role for gene transfer itself, pointing out that larger polyplexes lead to at least 10x increased gene transfer due to improved endosomal escape.

The excess of NPys during PEGylation in this surrounding could hinder oligomer crosslinking, theoretically leading to mono-functionalized oligomer-PEG constructs and not oligomer-oligomer constructs crosslinked by a bivalent linker. Although this mechanism is not clearly described, it was previously [242, 243] pointed out that pLL/pDNA polyplexes modified with pHMPA equipped with multiple attachment sites

mediated superior stability over linear (semitelechelic) pHMPA structures, considerably because of intermolecular cross-linking. However, for this shielding with enhanced stability, a slower reaction kinetic is described, suggesting that further experiments could be carried out after an enlonged PEGylation reaction with a lower amount of PEGylation reagent.

5 Summary

Non-viral gene delivery depicts a promising alternative to the classical, established method of viral gene delivery [1]. However, for a successful gene delivery by artificial vectors, these carriers need to display several functionalities. Besides the ability to bind and compact the cargo, they are also required to shade the nucleic acid from degradation. During formulation of the nanoparticles (so-called polyplexes), the particles need to exhibit a suitable size between 5.5 and 400 nm to accumulate within the tissue of interest [39, 40]. To tailor these vehicles to its needs, such as preventing the nucleic acid from degradation, or to reduce interaction with blood components [153], solid phase synthesis (SPS) was recently introduced [95, 98, 117, 131, 134].

Thereby, oligomers could be generated to draw clear-cut structure-activity relationships.

This first part of the study focuses on the optimization of shielding domains within sequence-defined oligomers (oligoaminoamides), which were generated by incorporation of novel artificial polyamino acids [131] for nucleic acid binding and compaction. Hereby, the strategy of pre-PEGylation, implying that a hydrophilic block of defined ethylene oxide repetitions was incorporated during oligomer synthesis, was applied to optimize pDNA delivery. To overcome the hampered transfection efficacy mediated by PEG [167], a hepatocyte growth factor receptor (HGFR) binding peptide (cmb) [181], that was previously found to mediate tumor-specific gene delivery in vitro as well as in vivo [64, 65], was introduced. Opposing requirements had to be dealt with while comparing shielding agents (PEG and peptidic Pro-Ala-Ser repeats) of different repetitions; both extremes (no shielding or very long shielding) have their drawbacks within two-arm oligoaminoamide pDNA polyplexes. However, it is concluded that a shorter shielding domain consisting of 12 PEG repetitions displays the best compromise for HGFR targeted gene delivery - this could also be shown by intratumoral delivery in vivo.

Besides pre-PEGylation, also the introduction of a targeted PEGylation reagent after polyplex formation facilitates tumor-specific gene delivery [169, 170]. This approach was explored within the second part of this thesis for EGFR (epidermal growth factor receptor) targeted pDNA polyplexes, complexed by a cysteine-containing

sequence-defined lipo-oligomer (454 [52]). Therefore, the EGFR specific peptide GE11 (YHWYGYTPQNVI) was linked to PEG of 24 repetitions and activated terminal cysteines (one or two respectively) which were assembled as mono- or bivalent PEGylation reagents via SPS. During biophysical evaluation and in vitro testing on different EGFR positive or low expressing tumor cells, pDNA lipopolyplexes post-modified with the bivalent PEGylation reagent via disulfide exchange chemistry, exhibited the highest EGFR dependent tumor uptake and gene expression, while circumventing polyplex aggregation observed for pre-PEGylated GE11 targeted 2-arm oligomers, which were evaluated side by side within this study.

Lipopolypexes are described to mediate increased polyplex stability, which is known to be of high importance for in vivo gene delivery. However, post-modified 454/pDNA polyplexes lack stability. Therefore new oligomers with increased cationic charge (mediated by increased Stp units from 4 to 6 or 8), improved hydrophobic stabilization (due to introduction of Y6 and/ or cholanic acid instead of oleic acid) as well as additional histidines (for endosomal buffering as well as potential stabilization via imidazole mediated - stacking [225, 235]) were introduced and then carefully evaluated via biophysical experiments to determine polyplex behavior in the presence of (poly)anionic stress and in the presence of serum. Here the novel oligomers 1175-1180 exhibited a significantly improved polyplex stability compared to 454. While in vitro transfections of the post-modified lipopolyplexes could not mediate gene delivery within 24 h, significant receptor-dependent cellular polyplex uptake was observed after 45 min for oligomers 1026, 1177 as well as 1176 and 1178 as well as 454.

Transfections of these post-modified polyplexes followed by addition of endosomolytic LPEI or chloroquine pointed out the functionality of this assay, as gene transfer could be significantly improved for 454/pDNA polyplexes. However, gene transfer was not notably enhanced for all other lipopolyplexes. This indicates that polyplexes formed with the newly generated lipo-oligomers might retain their cargo in subsequent steps preventing transformation of the pDNA towards its protein. Future in vivo experiments could show if stability was sufficiently improved to resist effects occurring in vivo and lipopolyplexes can overcome the intracellular barriers towards successful gene delivery.

6 Appendix